I updated indi with the most recent nightly, and my QHY163m won't connect properly. I opened a issue in github for this, but I thought I'd mention it here to see if others have seen this the problem. qhy_ccd_test seems to run fine with no errors, but connecting with EKOS as the client fails.

Logs attached.

jmh

PS: I have a 5 year old LCD monitor and I can't read the two instruction lines under the Attachment in the message composition pane because the contrast is terrible - light yellow txt on light background. Can we have light text on dark background or dark text on light background? Maximize contrast! This problem is common now on many other websites, and whoever started this new "style" guide for contrast needs to have his fingers removed from any keyboard...

see Constrast Rebellion for more info!

jmh

this is with libqhy SDK 5.0.1?

Yes.

More information on this:

1. I confirmed it is not the camera - it works with libqhy4.x on MacOS with no issues.
2. It also worked for a session on March 30 and this was with 5.0.1. Something changed between the 30th and April 3. All the cables and USB hubs appear to be fine as well.

jmh

I did a little more digging - after attempting to connect to the camera, I noticed that the pid changed. indi.output has the following debug message:

2019-04-05T02:50:17: Driver indi_qhy_ccd: terminate called after throwing an instance of 'std::bad_alloc'
2019-04-05T02:50:17: Driver indi_qhy_ccd: what(): std::bad_alloc

(Line 35,36 in the attached log.)

jmh

I informed QHY about this.

Btw, if the latest driver is linked against libqhy 4.X.X libs, it works fine? Just trying to eliminate any possibility with the driver being the issue.

I linked against 4.0.16 and it did not work - same behaviour.

On March 29, I pulled the latest bits from github, including 5.0.1 and everything worked for sessions on March 29 and 30. I did the same on April 3, and it stopped working.

I attached my QHY5L-II and it does work, although it won't connect automatically. The QHY163m has the built-in filter wheel, so perhaps the changes related to the filter wheel have some unintended consequences.

Regards,

jmh

Yes! QHY mentioned an issue with filter wheel causing a possible crash! I'll update them about this.

I'm not a git expert - how can I get the bits for a specific nightly version, say, current to March 29? If I can get those bits running, that should confirm the problem lies with the FW code somewhere.

jmh

actually, I just committed a possible fix for this, please try to pull from git and build the latest version now. or wait for nightly later.

I pulled down the latest and tried it:

1. no change in behaviour - logs attached.

2. indi_qhy_ccd died and was restarted - output from indiserver below:

2019-04-05T15:31:04: Driver indi_qhy_ccd: terminate called after throwing an instance of 'std::bad_alloc'
2019-04-05T15:31:04: Driver indi_qhy_ccd: what(): std::bad_alloc
2019-04-05T15:31:04: Driver indi_qhy_ccd: stderr EOF
2019-04-05T15:31:04: Driver indi_qhy_ccd: restart #1
Child process 11041 died

Run the driver via gdb in the terminal:
In Ekos, create a new profile, turn off Auto Connect and add Telescope Simulator + QHY CCD. Set mode to remote and leave it to localhost. In Logs, make sure to select Verbose Logging to File, check INDI + CCD. Click Start then in INDI Control Panel --> QHY CCD --> Options, make sure "Driver Debug" is checked. Then go back to Main Control Panel and press "Connect". Check the terminal to see if the driver crashed, if it did, type backtrace and paste the trace here. Also attach the KStars log file.